Predicting Well Inflow Using Computational Fluid Dynamics - Closer To The Truth?

Abstract

Abstract One of the great challenges facing our industry is accurate prediction of well inflow. Conventional methods have been too cumbersome and imprecise and have suffered from a lack of accuracy and clarity. Informed use of computational fluid dynamics, fine scale modelling and improved computing power enable far more accurate prediction of the impact of formation damage and thus of well performance. The accurate prediction of well performance helps with appraisal of development prospects, well planning and reliable prediction of true well and field value. If we know what the outcome of our actions and of our well designs will be, then we can make sensible and informed choices on damage impact and mitigation. Results from laboratory simulations of drilling and completion operations were generated from "standard" return permeability testing. The detailed data obtained, and its millimetre scale resolution was incorporated in to a well specific model. The damaged and undamaged states were examined using flow rate predictive computational fluid dynamics. The impact on flow in a single and dual permeability reservoir interval were calculated. In a specific example presented, the case for underbalance drilling was clearly made as the impact of overbalance drilling was predicted to have a severe impact on well productivity. The results of accurate and detailed laboratory simulation of formation damage have been translated using innovative software applications to give a prediction of well performance. This is the first time that computational fluid dynamics has been employed to predict well performance based on high quality laboratory testing. In future the laboratory tests will be designed to yield data most useful for the model and the model and grid scale will continue to be adjusted based on the specific challenge and objective. The detailed workflow used to achieve more accurate well performance prediction will be outlined in the paper. Introduction Modeling of permeability, pressure drops and formation damage in the near wellbore is an art that has suffered from lack of input and lack of tools in the past. The input deficit has largely been driven by poor or incomplete measurement and knowledge of formation damage mechanisms. These are not very often identified or quantified and some of the basic assumptions and rules of thumb for damage are simply incorrect. Even when properly understood and measured, existing industry software, inflow performance relationship models are inadequate tools to capture the detail and complexity of damage distribution. They generally accept only a very simple input of damage magnitude and thickness. Often data is altered in order to match the model. This is entirely the wrong way around. In building the workflows to model and understand damage, our guiding principal was that the model must alter to capture the damage rather than the other way around. Previous authors have attempted to model damage and create near wellbore inflow models which incorporate flow restrictions to a fine scale (Bennion et al 1996, Burton et al 1997, Han et al 2005, Yildiz 2004, Qutob and Ferreira 2005). These could be termed very near wellbore models of damage, most of which are numerical 2D models. Our objective here is to create a process through which complex very near and very, very near well inflow models can be created in 2D, 3D and 4D. The overriding objective is accurate prediction of well productivity and injectivity which will enable thorough well and completion designs and enable focus on real production barriers rather than speculation and heresay.